Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | "Duggan, Catherine R" <cduggan@fhcrc.org> |
| To | <statalist@hsphsun2.harvard.edu> |
| Subject | st: Cox regression- non linear trends |
| Date | Mon, 25 Jul 2011 18:27:19 -0700 |
Hello I have a Cox model examining Variable A and it's inhibitor Var B. Var. A seems to exhibit a non-linear trend when I examined the tertiles using a fully adjusted model, with mortality as an endpoint. The P.H. assumption was met. 1.00 ref 0.48 (95% CI 0.18-1.27) 0.97 (95% CI 0.35 - 2.67) (Though Var B was linear). Because of this, I thought that I should present just the categorical variable in the model. However my experience has been that reviewers always want to see both continuous and categorical variables presented, especially as I will present Var B both ways. Can anyone give me some advice please on how to handle non-linear continuous variables in Cox models? Many thanks Catherine * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/